Indoor Positioning via Particle Filtering Step Length Adaptation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing indoor positioning methods based on electromagnetic signals face challenges such as high error rates due to environmental factors and the need for accurate step length measurement, which affects positioning accuracy and efficiency.
Innovation Solution
A mobile terminal positioning method that uses a particle filtering algorithm to randomly allocate step lengths and update them based on usability ratings, reducing computation load and improving accuracy by reserving high-rated step lengths and eliminating low-rated ones, while also initializing new particles to maintain positioning efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the fingerprint database is drawn with longer signal acquisition time and more sampling points, then the positioning accuracy is improved, but the time costs and other expenses are higher
Solution Approach 1:
The patent performs fingerprint database drawing in advance before actual positioning operations. By pre-collecting signal strength data at multiple sampling points and storing it in a database, the system eliminates the need for time-consuming data collection during real-time positioning, thus resolving the contradiction between accuracy and time cost
Solution Approach 2:
The patent dynamically adjusts the number of sampling points and acquisition time based on the specific positioning scenario. Rather than using fixed parameters, the system can adaptively select appropriate sampling density and time duration, optimizing the balance between positioning accuracy and time consumption
2Ease of operation
If empirical values of step length are adopted to compute position changes, then the positioning process is simplified, but great position update errors occur due to user and environment differences
Solution Approach 1:
The patent uses particle filtering algorithms that continuously update particle weights based on how well they match observed signal strength data. This feedback mechanism allows the system to correct positioning estimates in real-time, compensating for errors caused by varying step lengths without requiring complex manual calibration
Solution Approach 2:
The patent changes the approach from using fixed empirical step length values to using probability distributions of step lengths. By representing step length as a random variable with a distribution rather than a fixed value, the system can account for user and environment variations while maintaining computational simplicity
3Device complexity
If RSSI is used for triangular positioning, then the positioning method is simple, but the error rate is great due to multipath effect
Solution Approach 1:
The patent introduces fingerprint databases and particle filtering as intermediary layers between the raw RSSI measurements and the final positioning result. These intermediaries process and filter the noisy RSSI data, reducing the impact of multipath effects while maintaining the simplicity of using signal strength for positioning
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Example positioning techniques include: generating an initial particle set, randomly allocating different moving step lengths for each particle in the initial particle set, updating position information of each particle at a previous moment to obtain a current particle set, rating usability of each particle according to position information of each particle in the current particle set and a signal fingerprint received at a current moment, acquiring a moving step length of each particle having a rating higher than a first predetermined threshold in the current particle set, and updating a moving step length of each particle having a rating lower than the first predetermined threshold according to the acquired moving step length. The techniques of the present disclosure obtain a moving step length closest to the actual step length of a person, and update the moving step length in time with the change of the step length of the person.